Cardiac organoids: novel regenerative therapy for myocardial infarction

Abstract Introduction Cell therapy for myocardial infarction (MI) faces two key challenges: poor cellular integration, requiring high therapeutic doses, and the risk of post-transplant arrhythmias. This project aimed to produce human cardiac organoids (COs) at a scalable manner, and validate their safety and efficacy as a regenerative therapy in a preclinical swine model of acute MI. Methods COs were derived from human induced pluripotent stem cells in stirred-tank bioreactors through self-aggregation into embryoid bodies followed by cardiac differentiation in a single workflow. A 100 mL protocol was scaled to 3 L, assessing process robustness and CO characteristics. Twelve immunosuppressed pigs underwent MI induction via coronary artery ligation and were randomized into Short-term CO (8-day follow-up, n=2), Long-term Control (30-day follow-up, n=4), and Long-term CO (30-day follow-up, n=6). Thirty minutes post-MI, treated animals received 3500 COs by intramyocardial injections, whereas controls received vehicle injections. Engraftment was assessed by immunohistofluorescence (IF) against human nuclear antigen (HNA) and spatial transcriptomics (n=1). In long-erm groups, rhythm was monitored for 15 days using ECG Holter recording; cardiac function and scar size were evaluated by magnetic resonance (MRI) at days 2 and 29; and electrophysiologic properties and arrhythmia inducibility at day 30 were assessed by high-density mapping (HDM) and programmed electrical stimulation, respectively. Results Bioreactor scaling increased output from 7000 COs/100 mL to 50000 COs/3 L while maintaining uniform size and organization, characterized by an endothelial cell outer layer and a cardiomyocyte-rich core, together with other cardiac cell types. IF analysis confirmed the presence of HNA+ cells in scar and peri-infarct zones in all studied follow up. At 8 days post-transplantation COs still maintained their spherical shape, whereas at 30 days COs were disaggregated, and cells migrated into myocardial scar. Spatial transcriptomics indicated reduced profibrotic gene expression in HNA+ cells at day 30 compared with day 8. All animals maintained a sinus rhythm without severe ventricular arrhythmias, and HDM and programmed electrical stimulation showed no group differences. MRI revealed significant increases in left ventricular (LV) stroke volume (p=0.05), LV ejection fraction (p=0.034), cardiac output (p=0.010), and cardiac index (p=0.048), along with smaller scar size (p=0.029) in the Long-term CO group, whereas these changes were not observed in controls. Conclusions Our bioreactor platform enables scalable, reproducible CO production. Following implantation, COs integrated robustly into the host myocardium, maintained long-term viability, and did not induce severe arrhythmias. Early findings pointed toward significant improvements in cardiac function and scar reduction, and larger ongoing studies are expected to further substantiate their therapeutic potential.